Tree Growth History, Stand Structure and Biomass of Premontane Forest at Cerro Tambo 35

Tree Growth History, Stand Structure, and Biomass of Premontane Forest Types
at the Cerro Tambo, Alto Mayo, Northern Peru

Historia del Crecimiento Arbóreo, Estructura Vertical, y Biomasa de Diferentes

Tipos de Bosque Premontanos en el Cerro Tambo, Alto Mayo, al Norte del Perú,

Reiner Zimmermann1,2*, Henry Soplin Roque3,4, Annett Börner1,2, Tobias Mette1,5

1) Ecological Botanical Gardens, Forest Ecology and Remote Sensing Group, University of
Bayreuth, Germany.
2) Max-Planck-Institute for Biogeochemistry, PO.Box 10 01 64, D-07701 Jena, Germany.
3) Universidad Nacional San Martin, Faculdad de Ecologia, Moyobamba, Peru
4) Manejo Ambiental, Proyecto Especial Alto Mayo, Moyobamba, Peru
5) Deutsche Luft- und Raumfahrtgesellschaft, Institut fuer Hochfrequenztechnik,
Oberpfaffenhofen, Germany
*corresponding authort: reiner.zimmermann@bgc-jena.mpg.de

Abstract
Tree growth and biomass accumulation were studied for two structurally contrasting premontane
forest types occuring from 1200 -1600 m a.s.l. at the Cerro Tambo, Region Alto Mayo in Peru.
The premontane vegetation at the Cerro Tambo consists of a mosaic of poor heath forests and
well developed premontane rain forests. Within each forest type, the variation in species
composition and aboveground biomass is small compared to the dramatic differences between
the heath forests and the rain forests. Mosaic type heath forest establishment in a zone with
potential dense rain forest cover could be a result of rare recurring natural fires on a secular
scale. This may trigger a series of successional stages which leads to irreversible deterioration of
site quality under current conditions. In order to test this hypothesis, information on stand
structure, tree growth rates, and tree age was obtained for describing the growth dynamics of the
contrasting forests and for detecting presence or absence of stand growth depression within the
life span of old growth individuals. Tree ring analysis of 106 trees from sites along an altitudinal
gradient showed that all forests of the Cerro Tambo area are characterized by low annual tree
growth rates. Heath forests and tall rain forests show pronounced difference in growth rates.
Heath forests are young and no individuals older than 45 years were found. In contrast, tall rain
forests trees of more than 150 years in age were found. We conclude that heath forests are most
likely successional stages after fairly recent disturbance, probably by natural fires. A linear
positive increase in tree growth in all forest plots and a further increase in the past two decades
may be attributed to a combined effect of the increasing atmospheric CO2 concentration and a
recent increase in atmospheric deposition of nutrients.

Resumen
Las transiciones abruptas entre el bosque lluvioso de alta montaña a lo largo grandes islas de
chamizales bajo condiciones locales idénticas de geología, topografía y clima es común en el
área del Cerro Tambo. La alteración por los derrumbes es insignificante. La influencia de la
presencia humana puede excluirse debido a la lejanía. Los chamizales tipo mosaico establecidos
cerca a una zona cubierta con potencial bosque lluvioso denso podrían ser el resultado de la
alteración natural. La única alteración creíble en esta área podría ser los incendios. Ninguna
evidencia de fuego directa se ha encontrado (carbón de leña), pero la vegetación natural puede
arder fácilmente y una observación en las áreas intervenidas muestra que los incendios se
provocan por sí mismos durante la estación seca. La frecuencia de incendios naturales podría ser
baja y en una escala esporádica. Los incendios también pueden activar una serie de fases

Lyonia 3(2): 291-300, 2003

36 Zimmermann, R., Soplin Roque, H., Börner, A. & T. Mette

sucesorias que llevan a la deterioración irreversible de la calidad del lugar bajo las condiciones
actuales. Bien podría ser, que estos eventos, - aún esporádicos -, lleven a cambios sustanciales en
las propiedades del suelo, en el rebrote de árboles y la recuperación de estos bosques lluviosos
altos puede tomar de décadas a siglos. Todos los bosques del área del Cerro Tambo, se
caracterizan por las proporciones anuales de crecimiento arbóreo muy bajas. Los chamizales y
los bosques lluviosos altos no muestran ninguna diferencia en las proporciones de crecimiento.
Los chamizales son jóvenes y ningún especímen mayor a 45 años fue encontrado. En contraste,
en los bosques lluviosos altos se encontraron árboles de más de 150 años de edad. Concluimos
que los chamizales pueden ser una fase sucesional después de una reciente alteración, lo más
probable los incendios naturales.

Introduction

Abrupt transitions from tall rain forest to large islands of low heath forest under identical local
conditions of geology, topography, and climate are commonly found in the Cerro Tambo Region
of the Alto Mayo Valley, Northern Peru (Fig. 1).

Fig.2. Tall rain forest on the flanks and heath forests on

Fig.1. Heath forest islands (blue) within tall top of a mountain ridge near the Cerro Tambo peak, Peru.
montane rain forests (magenta). The photo is taken at 1170 m a.s.l. (5.7216°S / 77.2604°
Forest map from Dempewolf (2000). W) 26.03.99

In this tropical montane area over cretaceous sandstone formations with moderate topography,
disturbance by landslides is very marginal and restricted to steep canyon walls (Fig.2). Recent
and historic anthropogenous influence can be excluded due to the remote location. Historically
the lowland terraces of the Alto Mayo were inhabited by a few Aguaruna tribes but the montane
region was visited for occasional hunting only. The sparse macro-fauna has no impact on the
forest distribution.

The differences in forest structure at the Cerro Tambo are quite striking. While the rain forest
vegetation have a closed canopy with tree heights of more than 30 m, the heath forests are shrubs
or small trees of up to 4 (max. 8 m) with grasses and herbs in the understorey. Heath forests
exhibit both xeromorphic structures as well as peinomorphic features, both supposedly pointing
toward increased competition for water and nutrient resources.

A detailed study of soils (Dempewolf, 2000; Börner 2001) showed that variation in soil structure
and composition could not explain the current vegetation distribution. In a subsequent eco-
physiologically oriented study of the vegetation water use and mineral nutrition by Mette (2001)
it turned out, that both vegetation types did neither differ much in water consumption, which was
low, nor in the nutritional status of plants which was also low but not critical. However it was

Lyonia 3(2): 291-300, 2003

Tree Growth History, Stand Structure and Biomass of Premontane Forest at Cerro Tambo 37

found that under conditions of an extended dry season, the rain forest was more susceptible to
water stress and drought damage and that soil water storage was more critical. During the study
of Mette (2001), the seasonal drought broke just a few days
before the soil water storage in the rain forest became
completely depleted. However, in drier years, the critical
point of irreversible drought damage could be reached. It was
observed that during the dry season man made slow burning
forest fires maintain themselves for several days until the next
rainfall, demonstrating the potential for spatially limited fire
disturbance in these montane rain forests.

The question remained open, what the cause of the peculiar

distribution pattern of both forest types was. It was suspected
Fig. 3. Tree disc of Ocotea sp. that natural disturbances must cause the occurance of heath
(Lauraceae), 157 years old, from a forests and that the disturbances must occur spatially
tall rain forest site at the Cerro irregular. Nothing could be said at this point about temporal
Tambo. patterns of disturbance, since the age of the forest types was
not known and successional stages were not obvious due to
the lack of transitional stages. Since fire and exceptional drought are both known disturbance
factors in forests, we investigated the climatic history, the stand and tree age and the growth
patterns over time of both vegetation types.

It was hypothesized, that

(1) Climatic data must indicate coincidence of extended dry seasons at maximum stand lifetime
for existing heath forests if drought matters
(2) Spatially limited but massive disturbance like fire must be reflected in stands by contemporal
tree establishment (similar maximum age of older individuals).
(3) Growth patterns in old stands must show systematic differences between stand type if
edaphic preconditions matter (site quality)

Material and Methods

Climate data from long term weather stations in the Alto Mayo valley and four own weather
stations along an altitudinal gradient were analysed. Forest structure, forest composition and soil
properties were studied on sites along an altitudinal gradient for heath forest stands and tall rain
forest stands. Tree age was analyzed at cross-sectional discs of 106 trees. Annual tree ring
formation (Worbes 1989) is present in most woody species of the area. This is most likely caused
by the pronounced seasonality in precipitation (Fig. 4). Tree growth ceases during the dry season
(Detienne 1989). Sesonal tree ring formation was confirmed in a blind test on discs of several
tree species from a helicopter landing site in the area which had a confirmed age of 31 years and
was abandoned after a one day use for setting up a trigonometric point in the forest.
Rioja
21.4
250
D ¡C
200
150
100
20.0
17.5 (880m)
avrg temperature
airtemperature
22.5
precipitation
50
precipitation
M
F
J
A40
15.0
JS
temperature
N
O
30
1964-1993 [mm]
[¡C] 1525 mm
Climate

The monthly precipitation sum in the Alto Mayo region may

fall to critical levels below 100 mm month-1 in some years.
Several extended periods of such conditions in the region were
recorded at two stations in the Alto Mayo valley over the past
fourty years with the most pronounced drought occuring in
1961 (Fig. 5). While tall rain forest and heath forests tolerate
short droughts, the vegetation and organic soil matter dries out
to critical levels during such times.

Lyonia 3(2): 291-300, 2003

Fig. 4. Climate diagram of

Rioja (880 m) in the Alto Mayo
38 Zimmermann, R., Soplin Roque, H., Börner, A. & T. Mette

400
300
200
0 Alto
Rioja
11960
00
Mopnthly
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984Mayo
Moyobamba Valley, Peru
precipitation (mm)

Fig. 5. Monthly sum of precipitation from 1959 to 1984 for two meteorological stations at Moyobamba
and Rioja in the Alto Mayo valley, Northern Peru.

Drought events at the eastern slopes of the Andes and the subsequent occurance of natural fires
could be due to El Niño events. El Niño events at the Pacific coast are supposed to lead to
drought effects in the following years along the Eastern Andes. The sea surface temperature of
the South Pacific tropical current as a proxy for El Niño events (ENE) in Peru increased slightly
from 1959 until recent with several positive deviations indication major ENE (Fig. 6a). Annual
precipitation in the Alto Mayo region varied from less than 1000 mm to 2400 mm over the same
time period but showed no clear temporal trend (Fig. 6b). Contrary to common assumption, no
correlation was found between ENE and years with low precipitation for the Alto Mayo valley.

5=
4
3
1
0South
-1
-2
r2 SD
mean
Deviation
Sea0.025
2500
1960
1970
1980
1990 Pacific
Surface
(deg Tropical Current
Temperature
C) Peru
=2000
Alto
2000
r2
1500
1000SD
mean
Annual
1960
1970
1980 Mayo
0.025
1990
2000 Valley,
Precipitation (mm)

Fig. 6a. Mean Sea surface temperature

deviation (SST) of the South Pacific
tropical current from 1958 until 1998.

Fig. 6 b. Annual precipitation in the

Alto Mayo valley, measured at
Moyobamba (data mising for 1984 and
1984 due to terrorist problems).

Lyonia 3(2): 291-300, 2003

Tree Growth History, Stand Structure and Biomass of Premontane Forest at Cerro Tambo 39

Study sites ranged along a transect with varying slopes but similar geology from 1202 to 1570 m
a.s.l. with four tall rain forests, two dense, and three open heath forests. Forest cover is much
lower in open heath forests. Basal area in one of the open heath forests was surprisingly high
with 27 m2 ha-1 (Table 1). Biomass of the tall rain forests (Ogawa et. Al.1965) is significantly
higher than in the heath forests and reaches 300 t ha-1 (Fig.7) while stem density is higher in most
heath forests. Based on the inventory data we suspected that either biomass accumulation (i.e.
growth) is lower in the heath forests than in the tall rain forest or that growth rates are
comparable but heath forests are significantly younger than tall rain forests and probably exposed
to more frequent and significant disturbance.

Stand age and growth rates

Mosaic type heath forests in a zone with potential dense rain forest cover could be the result of
natural disturbance. A plausible disturbance in this area are fires, but no direct fire evidence
(charcoal) has been found so far in soils. However, natural vegetation can be burnt easily and
observation in intervened areas shows that fires maintain themselves during the dry season. The
natural fire frequency could be low and on a secular scale. Fires may also trigger a series of
successional stages which lead to irreversible deterioration of site quality under current
conditions. It might well be, that these events, - while rare -, lead to substantial changes in the
soil properties for tree regrowth and the recovery to tall rain forest may take decades to centuries.

70
60
50
40
30 Dense Heath Forest, Cerro Tambo 50
40
30
20
10O pen Heath
IHumiria
D=120
ID=122
ID=123
Humiriaceae)
Normalized
0
Cumulative
1960
1970
1980
1990
2000 to Forest,
Chamizo
(Humiriaceae)
ID=121 balsamifera
year
stem Cerro
(
2000
increment Tambo
20
Normalized
1 01990
Cumulative
01960
1970
1980
2.5
2.0
1.5
1.02000
Annual
0.5
0.0 Forest ring
1960
1970
1980
1990
2000 to year
stem
type growth 2000Elevation
growth
(mm)
Plots (mm) Vegetation Slope % Basal area 2.5
2.0
1.5
1.0
0.5
Annual
0.0
1960
1970
1980
1990 350increment
2000 ring (mm) (mm) 10000
m.a.s.l cover at 1.2 m [m 2/ha] 300
-1
8000 -1
250
Tall rain forest T2 1412 92 52 51 200 6000
Tall rain forest T3 1570 92 0 _ 150 4000
Tall rain forest HLC06 1495 84 80 31 100 Stems ha
Tall rain forest HLC07 1437 94 36 22 Biomass t ha 2000
50
Dense heath forest T5 1406 82 26 17 0 0
Dense heath forest HLC04 1406 90 80 12
T all rain forest
Dense heathOpen heath
Open heath forest HLC05 1404 33 73 9
T2 forest forest
Open heath forest HLC09 1226 37 29 27
Open heath forest HLC10 1202 47 44 _
T5 HLC09

Table 1. Site and stand structure paramenters of Fig. 7. Biomass and stem density of seven
nine study sites in the Cerro Tambo region, study sites in the Cerro Tambo region, Peru.
Peru.

Fig. 8. Growth dynamics of trees in a dense heath forests (left) and a open heath forest (right) in the Cerro
Tambo region, Peru. Cumulative stem growth (top) and mean annual tree ring width (bottom) with standard
deviation (bars) are shown.

Analysis of tree growth and biomass accumulation showed that heath forests are young and all
trees on the open and dense heath forest plots turned out to be less than 40 years old (Fig.8). The
only tree individual which was found to be several years older showed a significant tree growth
depression between 1960 and 1962 followed by high increments in the next years. Growth rates
were then low, but between 1 and 1.5 mm year-1 with not much inter-annual variation. A
increasing tree growth trend over the last four decades was found in all heath forests plots.

Lyonia 3(2): 291-300, 2003

 40 Zimmermann, R., Soplin Roque, H., Börner, A. & T. Mette

Tall rain forest trees reach much higher age with a 157-year old Ocotea sp. (Fig. 3.) being the
oldest individual found from tree disk analysis. However, older individuals of probably twice this
age are likely but not yet confirmed since tree cores from much larger trees in the same area have
not yet been analysed for age. Tree age and annual growth varied more in tall rain forests
between individuals compared to heath forests. Mean growth rates for the tall rain forest stands
ranged from 0.5 to almost 2.0 mm year-1 and showed a larger inter-annual variability compared
to heath forests.

Tall rain forest growth at the Cerro Tambo shows variation, but no sign of major disturbances in
tree growth for the past 150 years. Temporal dynamics of tree growth rates for the last four
decades in the studied forest stands were not much different between heath forests and tall rain
forests. However, superimposed on a small positive linear trend in tree ring width increase over
the entire observation time of the past 150 years, an additional recent increase in tree growth in
all plots was found after ca. 1985. This increasing trend can be best demonstrated by the growth
patterns of the study plot with the longest record at the Cerro Tambo (Fig. 9).

We have no obvious explanation for this positive growth effect based on site properties or
climatic records in the region (cf. Fig. 6a and b). Since the tree ring width in all plots has not
been geometrically corrected for tree dimension actually a systematic decrease in ring with was
expected under steady growth conditions for older tree individuals. The increasing trend will be
even stronger if geometric correction will be applied.
744
2.0
1.5
Mean
1.0
0.5
stem
1920
1930
1940
1950
1960
1970
1980
1990
2000annual
increment (mm)

Fig. 9. Mean annual stem radius increment for trees of a tall rain forest site at 1570 m a.s.l. at the Cerro Tambo, Peru.
The time series ranges from 1920 until 2000 (n=9).

It seems less likely that only the increase in atmospheric CO2 promotes the observed growth
increase, since the increase of approx. 30 ppm in global CO2 concentration over the past 15 years
does physiologically not justify an observed >30 % increase in carbon allocation to stems. The
higher CO2 availability may be responsible for the positive linear long term trend observed.

For the recent positive shift in growth it is probable that an increase in nutrient deposition, either
by long distance transport or locally from the rural development activities in the neighbouring
Alto Mayo valley has lead to a better nutrient availability in these ecosystems. From soil analysis
it was clear that all forest types are stocking on poor soils and that especially the N and Mg, P, K
supply is low (Demewolf 200, Boerner 2001). Consequently, all stands will react positively to an
improvement in nutrient supply. The Carretera Marginal has opened the Alto Mayo valley in
1977 to traffic. Intensive agricultural activities, including the application of industrial fertilizers
for rice cultivation, started in the early eighties. The agricultural area of the Alto Mayo valley
and the Carretera Marginal are less than 20 km direct line away from the Cerro Tambo.
Tree growth rates in the Alto Mayo rgeion decrease with increasing altitude (Fig 11).The studied
tall montane rain forests and heath forests at the Cerro Tambo show a significantly lower growth

Lyonia 3(2): 291-300, 2003

Tree Growth History, Stand Structure and Biomass of Premontane Forest at Cerro Tambo 41

than forests of the lowlands. The hill areas of the Alto Mayo region vary in growth rates. This
variation is caused by different edaphic soil properties. In general, the higher thermal regime in
the lowlands and the significantly better nutrient supply on alluvial soils are responsible for high
growth rates.

Conclusions
2.4
Cerro Tambo,
2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
Annual
0.4 ringAlto
treename Mayo,
growth Peru
(mm)
0.2
0.0
T3
T2
H6&7
H8
H4&5
H10 Plot
Tree
10234571000
Annual innundation
hill
Cerro
50 region
Tambo forest
increment
Altitude
1250
1500
1750 [m a.s.l.] [mm/a]

Fig. 10. Annual mean radial tree ring growth for Fig. 11. Annual radial increment of trees of the
six study sites at the Cerro Tambo, Peru. Alto Mayo region, Peru.

All forests of the Cerro Tambo area are characterized by low annual tree growth rates. Heath
forests and tall rain forests show pronounced difference in growth rates. Heath forests are young
and no individuals older than 45 years were found. In contrast, tall rain forests trees of more than
150 years in age were found. We conclude that heath forests are most likely successional stages
after fairly recent disturbance, probably by natural fires. This needs to be confirmed by direct
evidence. A linear positive increase in tree growth in all forest plots and a further increase in the
past two decades may be attributed to a combined effect of the increasing atmospheric CO2
concentration and a recent increase in atmospheric deposition of nutrients.

Acknowledgements

We thank Arno Perisutti (GTZ-DIAM) in Moyobamba, Peru for his invaluable support during
our field work. Funding for this study was provided by the Gesellschaft für technische
Zusammenarbeit, Germany (GTZ). We thank the representatives of the Peruvian Government,
Proyecto Especial Alto Mayo (PEAM) and the local authorities of the Comuniades Nativas
Aguaruna for the permission to perform field work at the Cerro Tambo. We thank Juan Castro,
Eleaser Gamonal (both PEAM), Marianne Lauerer (ÖBG Bayreuth), Jan Dempewolf, Ruben
Ruiz and many students from the Universidad Nacional San Martin for substantial help during
the field work. We feel deeply indebted to the local people of the Cerro Tambo forest region for
their hospitality. Without them this study would not have been possible.

References

Boerner, A. 2001. Klassifikation submontaner tropischer Wälder am Andenostabhang im

Einzugsgebiet des Río Avisado, Region Alto Mayo, Nordperú Diplomarbeit,
Universität Bayreuth.

Lyonia 3(2): 291-300, 2003

42 Zimmermann, R., Soplin Roque, H., Börner, A. & T. Mette

Dempewolf, J. 2000. Classification of montane rain forests on the eastern slopes of the
Peruvian Andes, in the Río Avisado and Río Tioyacu watersheds. Diplomarbeit, Universität
Bayreuth.
Detienne P. 1989. Appearance and periodicity of growth rings in some tropical woods. IAWA
Bulletin 10(2): 123-132.
Mette T. 2001. Water use strategies of dominant tree species in montane forest biomes of
the Cerro Tambo, Northern Peru. Diplomarbeit, Universität Bayreuth.
Ogawa, H., K. Yoda, K. Ogino, & T. Kira. 1965. Compatarive ecological studies on three
main types of forest vegetation in Thailand, II. Plant Biomass. Nature and Life in
Southeast Asia, 4: 49-80
Worbes M. 1989. Growth rings, increment and age of trees in forests, savannas and mountain
forests in the Neotropics. IAWA Bulletin 10(2): 109-122.

Lyonia 3(2): 291-300, 2003

Tree Growth History, Stand Structure and Biomass of Premontane Forest at Cerro Tambo 43

Lyonia 3(2): 291-300, 2003